1 Billion Row Challenge... in Ruby?

1_trivial.rb

stats = {}

File.open('measurements.txt') do |f|
  f.each_line do |line|
    city, temp_s = line.split(';')
    temp = temp_s.to_f
    stats[city] ||= {
      min: 10_000,
      max: -10_000,
      sum: 0,
      count: 0
    }
    stats[city][:min] = temp if temp < stats[city][:min]
    stats[city][:max] = temp if temp > stats[city][:max]
    stats[city][:sum] += temp
    stats[city][:count] += 1
  end
end

stats.each do |city, cstats|
  puts "#{city}=#{cstats[:min]}/#{cstats[:sum].to_f / cstats[:count]}/#{cstats[:max]}"
end

9_final.rb

stats = Hash.new { |h, k| h[k] = [10_000, -10_000, 0, 0] }

SEMI = String.new(';', encoding: Encoding::BINARY)
NEWLINE = String.new("\n", encoding: Encoding::BINARY)

File.open('measurements.txt', encoding: Encoding::BINARY) do |f|
  loop do
    city = f.readline(SEMI)
    temp = f.readline(NEWLINE).to_f
    existing = stats[city]
    min, max, sum, count = existing
    existing[0] = temp if temp < min
    existing[1] = temp if temp > max
    existing[2] = sum + temp
    existing[3] = count + 1
  end
rescue EOFError
  # all good
end

stats.each do |city, cstats|
  puts "#{city[0..-2]}=#{cstats[0]}/#{cstats[2] / cstats[3]}/#{cstats[1]}"
end